Method for optimizing redundant graphics

Abstract

The invention provides an algorithm for adding and optimizing a redundant graph. The algorithm comprises the following steps of: searching a positive abnormal region and a negative abnormal region ofa version graph; optimizing the redundant graph of the positive abnormal area and optimizing the redundant graph of the negative abnormal area for the graph of the positive abnormal area; enabling thechemical mechanical planarization model to predict the surface topography of the optimized redundant graph; and if the predicted surface topography reaches a set value, successfully optimizing the redundant graph added in the layout. According to the method, the existing redundant graph is optimized, and the surface appearance of the redundant graph predicted and optimized by the chemical mechanical planarization model is used as a feedback means of the added redundant graph, so that the uniformity of local graph distribution of layouts of different layers among different products can be optimized, and the optimization effect of redundant graph addition is further improved; and the uniformity of the surface morphology of each layer after process machining is improved, and the process defects of the subsequent manufacturing process are reduced.

Description

technical field [0001] The invention relates to the technical field of semiconductors, in particular to a method for optimizing redundant graphics. Background technique [0002] In order to achieve a uniform distribution of integrated circuit layout graphics to improve the process window of related processes that depend on layout graphics distribution in the manufacturing process of integrated circuits, such as chemical mechanical planarization processes, it is usually necessary to add redundant graphics in areas where integrated circuit layout graphics are sparsely distributed To make the distribution of layout graphics relatively uniform, thereby reducing process defects in the subsequent manufacturing process and improving product yield. [0003] The existing redundant graphics addition is usually based on fixed redundant graphics geometric features, that is, different product layouts are added using the same redundant graphics structure and addition rules under the same ...

AI Technical Summary

Problems solved by technology

[0003] The existing redundant graphics addition is usually based on fixed redundant graphics geometric features, that is, different product layouts are added using the same redundant graphics structure and addition rules under the same process conditions, which can make layout graphics global The density is relatively uniform, but because the initial graphic distribution of different products is different, when using the unified redundant graphic addition rule to add, the local graphic distribution of different products cannot be optimal at the same time, and other geometric features such as graphics Line width, graphic perimeter, etc. are not included in the optimization direction

In addition, the traditional inspection of the added layout graphics only checks the density distribution of the layout graphics, the geometric characteristics of the graphics themselves, and the relationship between graphics

These inspections do not take into account the influence of the underlying graphics on the surface topography of the upper layer, that is, the lamination effect is not considered, and cannot directly reflect the exact surface topography changes after processing at each level

Images